SIVA-RL: Sensitivity-Invariance Visual Alignment for Multimodal Reinforcement Learning
Reinforcement learning with verifiable rewards (RLVR) drives multimodal reasoning, but answer-level correctness does not guarantee that a vision-language model grounds its predictions in visual evidence. Existing visual-intervention methods contrast policy behavior on original and modified images, yet assign supervision by the type of intervention rather than its observed effect. This assumption fails: identical operators produce heterogeneous outcomes across samples. We propose SIVA-RL, a Sensitivity-Invariance Visual Alignment framework that replaces operator-conditioned regularization with
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- FuzzySimilar title/name (fuzzy) · 59%aymericdamien/TopDeepLearning →
“Fuzzy title match (0.73): “SIVA-RL: Sensitivity-Invariance Visual Alignment for Multimo” ≈ “aymericdamien/TopDeepLearning””
- LinkedLinked via arxiv author · 85%Jincheng Tang →
“SIVA-RL: Sensitivity-Invariance Visual Alignment for Multimodal Reinforcement Learning”
- LinkedLinked via arxiv author · 85%Junzhi Ning →
“SIVA-RL: Sensitivity-Invariance Visual Alignment for Multimodal Reinforcement Learning”
- LinkedLinked via arxiv author · 85%Min Cen →
“SIVA-RL: Sensitivity-Invariance Visual Alignment for Multimodal Reinforcement Learning”
- LinkedLinked via arxiv author · 85%Shiwei Lin →
“SIVA-RL: Sensitivity-Invariance Visual Alignment for Multimodal Reinforcement Learning”
- LinkedLinked via arxiv author · 85%Xinyi Zeng →
“SIVA-RL: Sensitivity-Invariance Visual Alignment for Multimodal Reinforcement Learning”
- LinkedLinked via arxiv author · 85%Pinxian Zeng →
“SIVA-RL: Sensitivity-Invariance Visual Alignment for Multimodal Reinforcement Learning”
- LinkedLinked via arxiv author · 85%Rongbin Li →
“SIVA-RL: Sensitivity-Invariance Visual Alignment for Multimodal Reinforcement Learning”